Clustered call park issues

[acejavelin]

I have a customer that has an 11 site cluster, with resilient phones across various nodes, that was just upgraded from 9.0 to MCD 4.2 (all the farther we could go due to other equipment compatibility).

Here is the issue I am having:

Call comes in to Node 2 on a PRI and is answered by a phone that is resilient to Node 2&1 (Node 2 primary, 1 secondary), and parks the call on a resilient phantom hunt group extension (Node 1 primary, node 2 secondary). When this happen the caller hears MOH for about 3 seconds then silence, but is still parked and can be retrieved. If the user just puts the call on hold, or parks it on an actual extension, the MOH works fine.

If a call comes in to Node 1's PRI and is answered by a phone that is resilient to Node 1&2 and parks it on the same hunt group, the MOH works fine.

I never heard a complaint of this issue on 9.0.

The only thing I can see in programming is that is COS option "Local Music on Hold source" on the PRI's COS, I have tried it both ways with no change in the behavior... What am I missing?

[ralph]

The 3 seconds strikes me as bizarre.   I can't think of anything that would cause that.
Wondering if there is some type of camp on going on.
Try giving the hunt group it's own COS then playing with the timers in COS.
Don't know what to look for but it may find something in there to give you a clue.

Ralph

[acejavelin]

Has something to do with the clustering, I changed the COS of the IP Trunks (added the Local Music on Hold) and now I get 2-3 seconds of MOH, then it jumps to a different spot in the file which I am assuming is jumping controllers in the cluster... weird... and not what the documentation suggests, it only talks about MOH being supplied by the parker or the caller's cluster element, not the park location's element. Gave up playing with it yesterday, gonna hit it fresh this morning.

Weird part is, this never came up after years of running 9.0, but did with a couple weeks of upgrading to 4.2